GitHub is letting developers choose between Copilot and its biggest rivals
GitHub subscribers now have a choice of coding agents to help them create.
In addition to GitHub’s own Copilot, users can choose from Anthropic’s Claude Code, OpenAI’s Codex, and custom agents on GitHub and in VS Code, through a new feature GitHub calls Agent HQ.
AgentHQ is currently available to Copilot Pro+ and Enterprise subscribers, but the company plans to bring access to Claude and Codex to other Copilot subscription tiers soon.
The company first announced this new capability at its GitHub Universe event in October 2025, with plans to launch it later that year (though that timeline clearly shifted a bit).
All of these software development agents already integrate with and use GitHub in some form. But with Agent HQ, developers can now work with all of these agents from a consolidated dashboard — and they can switch between them as needed for different tasks as well.
Another interesting use case here is assigning the same task to different agents to see how they reason over a given problem and which solutions they may offer.
As GitHub COO Kyle Daigle told The New Stack during Agent HQ’s announcement last year, the company wants to give developers choice, whether that’s which IDE they want to use or which agent they prefer to work with.
“Our goal with Agent HQ,” Daigle said at the time, “is that we have a single place where you can use basically any coding agent that wants to integrate, and have a single pane of glass — a mission control interface, where I can see all the tasks, what they’re doing, what state of code they’re in — think creation, code review, etc, and offer up the underlying primitives that have let us build GitHub’s Copilot coding agent to all of those other coding agents.”
Model choice in GitHub (credit: GitHub).
Developers can assign any task to their preferred agent and let it work asynchronously. Claude and Codex sessions can be started from GitHub.com, the GitHub mobile app, and in VS Code. Developers can assign issues to these agents, and the agents can submit draft pull requests for review.
Agents can also work on existing pull requests, and as has become standard, mentioning @Copilot, @Claude, or @Codex in PR comments will also kick off any follow-up work.
GitHub notes that all of the work these agents do will be logged for future reviews, “so their output fits naturally into the same workflows you already use to evaluate developer contributions.”
Each session will consume one premium request. Those premium requests are GitHub’s currency for using more advanced features and premium models. Copilot Pro+ users get 1,500 of those requests, and Enterprise users get 1,000, at a cost of $0.04 per additional request.
